ml_model_visualization.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Install required libraries\n",
    "!pip install matplotlib thop torchviz torchview torchvision torchsummary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "from torchsummary import summary\n",
    "from torchviz import make_dot\n",
    "from torchview import draw_graph\n",
    "from IPython.display import Image, display\n",
    "from thop import profile\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "NeuralArchitect(\n",
      "  (convolutional_layer): Conv2d(3, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
      "  (activation): ReLU()\n",
      "  (dense_layer1): Linear(in_features=12544, out_features=128, bias=True)\n",
      "  (dense_layer2): Linear(in_features=128, out_features=10, bias=True)\n",
      ")\n",
      "Output shape: torch.Size([1, 10])\n"
     ]
    }
   ],
   "source": [
    "class NeuralArchitect(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(NeuralArchitect, self).__init__()\n",
    "        self.convolutional_layer = nn.Conv2d(3, 16, kernel_size=3, stride=1, padding=1)\n",
    "        self.activation = nn.ReLU()\n",
    "        self.dense_layer1 = nn.Linear(16 * 28 * 28, 128)\n",
    "        self.dense_layer2 = nn.Linear(128, 10)\n",
    "\n",
    "    def forward(self, input_tensor):\n",
    "        x = self.convolutional_layer(input_tensor)\n",
    "        x = self.activation(x)\n",
    "        x = x.view(x.size(0), -1)\n",
    "        x = self.dense_layer1(x)\n",
    "        x = self.activation(x)\n",
    "        x = self.dense_layer2(x)\n",
    "        return x\n",
    "\n",
    "# Create an instance of NeuralArchitect\n",
    "network = NeuralArchitect()\n",
    "\n",
    "# Create a sample input\n",
    "sample_input = torch.randn(1, 3, 28, 28)\n",
    "\n",
    "# Generate the output\n",
    "output = network(sample_input)  # Note: I used network(sample_input) instead of network.construct(sample_input)\n",
    "\n",
    "# Print the network structure\n",
    "print(network)\n",
    "\n",
    "# Print the output shape\n",
    "print(\"Output shape:\", output.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This method excels at basic layer visualization but falls short in conveying architectural complexity. It presents a simplified view, ideal for quick overviews, yet fails to capture intricate layer interactions and data flow. For an in-depth understanding, I'll employ more advanced techniques."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualize Neural Network Architecture with PyTorch and Graphviz\n",
    "\n",
    "This code snippet demonstrates the process of creating, visualizing, and analyzing a neural network using PyTorch and Graphviz. It generates a computational graph of the NeuralArchitect model, renders it as an image for visual inspection, and provides essential information about the network's structure and output shape."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "NeuralArchitect(\n",
      "  (convolutional_layer): Conv2d(3, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
      "  (activation): ReLU()\n",
      "  (dense_layer1): Linear(in_features=12544, out_features=128, bias=True)\n",
      "  (dense_layer2): Linear(in_features=128, out_features=10, bias=True)\n",
      ")\n",
      "Output shape: torch.Size([1, 10])\n"
     ]
    }
   ],
   "source": [
    "# Define NeuralArchitect class\n",
    "network = NeuralArchitect()\n",
    "\n",
    "# Create a sample input\n",
    "sample_input = torch.randn(1, 3, 28, 28)\n",
    "\n",
    "# Generate the output\n",
    "output = network(sample_input)  # Changed from network.construct to network\n",
    "\n",
    "# Create the graph\n",
    "graph = make_dot(output, params=dict(network.named_parameters()))\n",
    "\n",
    "# Set the path to the dot executable in the environment\n",
    "# Make sure to replace this with your actual dot path if different\n",
    "os.environ[\"PATH\"] += os.pathsep + \"/usr/local/bin/dot\"\n",
    "\n",
    "# Render the graph to a temporary file\n",
    "temp_filename = \"temp_neural_structure\"\n",
    "graph.render(temp_filename, format=\"png\", cleanup=True)\n",
    "\n",
    "# Display the image\n",
    "display(Image(filename=f\"{temp_filename}.png\"))\n",
    "\n",
    "# Remove the temporary file\n",
    "os.remove(f\"{temp_filename}.png\")\n",
    "\n",
    "# Print the network structure\n",
    "print(network)\n",
    "\n",
    "# Print the output shape\n",
    "print(\"Output shape:\", output.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "NeuralArchitect(\n",
      "  (convolutional_layer): Conv2d(3, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
      "  (activation): ReLU()\n",
      "  (dense_layer1): Linear(in_features=12544, out_features=128, bias=True)\n",
      "  (dense_layer2): Linear(in_features=128, out_features=10, bias=True)\n",
      ")\n",
      "----------------------------------------------------------------\n",
      "        Layer (type)               Output Shape         Param #\n",
      "================================================================\n",
      "            Conv2d-1           [-1, 16, 28, 28]             448\n",
      "              ReLU-2           [-1, 16, 28, 28]               0\n",
      "            Linear-3                  [-1, 128]       1,605,760\n",
      "              ReLU-4                  [-1, 128]               0\n",
      "            Linear-5                   [-1, 10]           1,290\n",
      "================================================================\n",
      "Total params: 1,607,498\n",
      "Trainable params: 1,607,498\n",
      "Non-trainable params: 0\n",
      "----------------------------------------------------------------\n",
      "Input size (MB): 0.01\n",
      "Forward/backward pass size (MB): 0.19\n",
      "Params size (MB): 6.13\n",
      "Estimated Total Size (MB): 6.33\n",
      "----------------------------------------------------------------\n",
      "Total parameters: 1607498\n",
      "Trainable parameters: 1607498\n",
      "Conv2d output shape: torch.Size([1, 16, 28, 28])\n",
      "ReLU output shape: torch.Size([1, 16, 28, 28])\n",
      "Linear output shape: torch.Size([1, 128])\n",
      "ReLU output shape: torch.Size([1, 128])\n",
      "Linear output shape: torch.Size([1, 10])\n",
      "NeuralArchitect output shape: torch.Size([1, 10])\n",
      "Conv2d output shape: torch.Size([1, 16, 28, 28])\n",
      "ReLU output shape: torch.Size([1, 16, 28, 28])\n",
      "Linear output shape: torch.Size([1, 128])\n",
      "ReLU output shape: torch.Size([1, 128])\n",
      "Linear output shape: torch.Size([1, 10])\n",
      "NeuralArchitect output shape: torch.Size([1, 10])\n",
      "convolutional_layer activation stats:\n",
      "  Mean: -0.0092\n",
      "  Std: 0.5674\n",
      "  Min: -2.3112\n",
      "  Max: 2.0532\n",
      "activation activation stats:\n",
      "  Mean: 0.0737\n",
      "  Std: 0.1153\n",
      "  Min: 0.0000\n",
      "  Max: 0.4537\n",
      "dense_layer1 activation stats:\n",
      "  Mean: -0.0337\n",
      "  Std: 0.2221\n",
      "  Min: -0.5984\n",
      "  Max: 0.4537\n",
      "dense_layer2 activation stats:\n",
      "  Mean: -0.0025\n",
      "  Std: 0.0787\n",
      "  Min: -0.1779\n",
      "  Max: 0.1388\n"
     ]
    }
   ],
   "source": [
    "# Create an instance of the NeuralArchitect\n",
    "network = NeuralArchitect()\n",
    "\n",
    "# Print the network architecture\n",
    "print(network)\n",
    "\n",
    "# Use torchsummary to get a detailed summary of the network\n",
    "summary(network, (3, 28, 28))\n",
    "\n",
    "# Analyse the network parameters\n",
    "total_params = sum(p.numel() for p in network.parameters())\n",
    "trainable_params = sum(p.numel() for p in network.parameters() if p.requires_grad)\n",
    "\n",
    "print(f\"Total parameters: {total_params}\")\n",
    "print(f\"Trainable parameters: {trainable_params}\")\n",
    "\n",
    "# Examine the output shape at each layer\n",
    "sample_input = torch.randn(1, 3, 28, 28)\n",
    "\n",
    "def hook_fn(module, input, output):\n",
    "    print(f\"{module.__class__.__name__} output shape: {output.shape}\")\n",
    "\n",
    "for name, module in network.named_modules():\n",
    "    if not isinstance(module, nn.Sequential):\n",
    "        module.register_forward_hook(hook_fn)\n",
    "\n",
    "_ = network(sample_input)\n",
    "\n",
    "# Analyse the activation distributions\n",
    "activations = {}\n",
    "\n",
    "def activation_hook(name):\n",
    "    def hook(module, input, output):\n",
    "        activations[name] = output.detach()\n",
    "    return hook\n",
    "\n",
    "for name, module in network.named_modules():\n",
    "    if isinstance(module, (nn.ReLU, nn.Linear, nn.Conv2d)):\n",
    "        module.register_forward_hook(activation_hook(name))\n",
    "\n",
    "_ = network(sample_input)\n",
    "\n",
    "for name, activation in activations.items():\n",
    "    print(f\"{name} activation stats:\")\n",
    "    print(f\"  Mean: {activation.mean().item():.4f}\")\n",
    "    print(f\"  Std: {activation.std().item():.4f}\")\n",
    "    print(f\"  Min: {activation.min().item():.4f}\")\n",
    "    print(f\"  Max: {activation.max().item():.4f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "NeuralArchitect(\n",
      "  (convolutional_layer): Conv2d(3, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
      "  (activation): ReLU()\n",
      "  (dense_layer1): Linear(in_features=12544, out_features=128, bias=True)\n",
      "  (dense_layer2): Linear(in_features=128, out_features=10, bias=True)\n",
      ")\n",
      "----------------------------------------------------------------\n",
      "        Layer (type)               Output Shape         Param #\n",
      "================================================================\n",
      "            Conv2d-1           [-1, 16, 28, 28]             448\n",
      "              ReLU-2           [-1, 16, 28, 28]               0\n",
      "            Linear-3                  [-1, 128]       1,605,760\n",
      "              ReLU-4                  [-1, 128]               0\n",
      "            Linear-5                   [-1, 10]           1,290\n",
      "================================================================\n",
      "Total params: 1,607,498\n",
      "Trainable params: 1,607,498\n",
      "Non-trainable params: 0\n",
      "----------------------------------------------------------------\n",
      "Input size (MB): 0.01\n",
      "Forward/backward pass size (MB): 0.19\n",
      "Params size (MB): 6.13\n",
      "Estimated Total Size (MB): 6.33\n",
      "----------------------------------------------------------------\n",
      "Total parameters: 1607498\n",
      "Trainable parameters: 1607498\n",
      "Conv2d output shape: torch.Size([1, 16, 28, 28])\n",
      "ReLU output shape: torch.Size([1, 16, 28, 28])\n",
      "Linear output shape: torch.Size([1, 128])\n",
      "ReLU output shape: torch.Size([1, 128])\n",
      "Linear output shape: torch.Size([1, 10])\n",
      "NeuralArchitect output shape: torch.Size([1, 10])\n",
      "Conv2d output shape: torch.Size([1, 16, 28, 28])\n",
      "ReLU output shape: torch.Size([1, 16, 28, 28])\n",
      "Linear output shape: torch.Size([1, 128])\n",
      "ReLU output shape: torch.Size([1, 128])\n",
      "Linear output shape: torch.Size([1, 10])\n",
      "NeuralArchitect output shape: torch.Size([1, 10])\n",
      "convolutional_layer activation stats:\n",
      "  Mean: -0.0431\n",
      "  Std: 0.5570\n",
      "  Min: -2.2946\n",
      "  Max: 2.2638\n",
      "activation activation stats:\n",
      "  Mean: 0.1091\n",
      "  Std: 0.1440\n",
      "  Min: 0.0000\n",
      "  Max: 0.6946\n",
      "dense_layer1 activation stats:\n",
      "  Mean: 0.0312\n",
      "  Std: 0.2310\n",
      "  Min: -0.5421\n",
      "  Max: 0.6946\n",
      "dense_layer2 activation stats:\n",
      "  Mean: 0.0217\n",
      "  Std: 0.0626\n",
      "  Min: -0.1269\n",
      "  Max: 0.0764\n",
      "Conv2d output shape: torch.Size([1, 16, 28, 28])\n",
      "ReLU output shape: torch.Size([1, 16, 28, 28])\n",
      "Linear output shape: torch.Size([1, 128])\n",
      "ReLU output shape: torch.Size([1, 128])\n",
      "Linear output shape: torch.Size([1, 10])\n",
      "NeuralArchitect output shape: torch.Size([1, 10])\n",
      "convolutional_layer.weight gradient stats:\n",
      "  Mean: 0.0103\n",
      "  Std: 0.1311\n",
      "  Min: -0.3861\n",
      "  Max: 0.4249\n",
      "convolutional_layer.bias gradient stats:\n",
      "  Mean: -0.0127\n",
      "  Std: 0.0860\n",
      "  Min: -0.1501\n",
      "  Max: 0.1272\n",
      "dense_layer1.weight gradient stats:\n",
      "  Mean: 0.0031\n",
      "  Std: 0.0446\n",
      "  Min: -0.8048\n",
      "  Max: 0.7880\n",
      "dense_layer1.bias gradient stats:\n",
      "  Mean: 0.0157\n",
      "  Std: 0.1197\n",
      "  Min: -0.3555\n",
      "  Max: 0.3481\n",
      "dense_layer2.weight gradient stats:\n",
      "  Mean: 0.1091\n",
      "  Std: 0.1434\n",
      "  Min: 0.0000\n",
      "  Max: 0.6946\n",
      "dense_layer2.bias gradient stats:\n",
      "  Mean: 1.0000\n",
      "  Std: 0.0000\n",
      "  Min: 1.0000\n",
      "  Max: 1.0000\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x1200 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Conv2d output shape: torch.Size([1, 16, 28, 28])\n",
      "\n",
      "Layer: Conv2d\n",
      "Input shape: torch.Size([1, 3, 28, 28])\n",
      "Output shape: torch.Size([1, 16, 28, 28])\n",
      "Parameters: 448\n",
      "Weight stats:\n",
      "  Mean: -0.0072\n",
      "  Std: 0.1088\n",
      "Bias stats:\n",
      "  Mean: -0.0412\n",
      "  Std: 0.0895\n",
      "ReLU output shape: torch.Size([1, 16, 28, 28])\n",
      "Linear output shape: torch.Size([1, 128])\n",
      "\n",
      "Layer: Linear\n",
      "Input shape: torch.Size([1, 12544])\n",
      "Output shape: torch.Size([1, 128])\n",
      "Parameters: 1605760\n",
      "Weight stats:\n",
      "  Mean: 0.0000\n",
      "  Std: 0.0052\n",
      "Bias stats:\n",
      "  Mean: -0.0005\n",
      "  Std: 0.0050\n",
      "ReLU output shape: torch.Size([1, 128])\n",
      "Linear output shape: torch.Size([1, 10])\n",
      "\n",
      "Layer: Linear\n",
      "Input shape: torch.Size([1, 128])\n",
      "Output shape: torch.Size([1, 10])\n",
      "Parameters: 1290\n",
      "Weight stats:\n",
      "  Mean: 0.0024\n",
      "  Std: 0.0519\n",
      "Bias stats:\n",
      "  Mean: 0.0015\n",
      "  Std: 0.0489\n",
      "NeuralArchitect output shape: torch.Size([1, 10])\n",
      "[INFO] Register count_convNd() for <class 'torch.nn.modules.conv.Conv2d'>.\n",
      "[INFO] Register zero_ops() for <class 'torch.nn.modules.activation.ReLU'>.\n",
      "[INFO] Register count_linear() for <class 'torch.nn.modules.linear.Linear'>.\n",
      "Conv2d output shape: torch.Size([1, 16, 28, 28])\n",
      "\n",
      "Layer: Conv2d\n",
      "Input shape: torch.Size([1, 3, 28, 28])\n",
      "Output shape: torch.Size([1, 16, 28, 28])\n",
      "Parameters: 448\n",
      "Weight stats:\n",
      "  Mean: -0.0072\n",
      "  Std: 0.1088\n",
      "Bias stats:\n",
      "  Mean: -0.0412\n",
      "  Std: 0.0895\n",
      "ReLU output shape: torch.Size([1, 16, 28, 28])\n",
      "Linear output shape: torch.Size([1, 128])\n",
      "\n",
      "Layer: Linear\n",
      "Input shape: torch.Size([1, 12544])\n",
      "Output shape: torch.Size([1, 128])\n",
      "Parameters: 1605760\n",
      "Weight stats:\n",
      "  Mean: 0.0000\n",
      "  Std: 0.0052\n",
      "Bias stats:\n",
      "  Mean: -0.0005\n",
      "  Std: 0.0050\n",
      "ReLU output shape: torch.Size([1, 128])\n",
      "Linear output shape: torch.Size([1, 10])\n",
      "\n",
      "Layer: Linear\n",
      "Input shape: torch.Size([1, 128])\n",
      "Output shape: torch.Size([1, 10])\n",
      "Parameters: 1290\n",
      "Weight stats:\n",
      "  Mean: 0.0024\n",
      "  Std: 0.0519\n",
      "Bias stats:\n",
      "  Mean: 0.0015\n",
      "  Std: 0.0489\n",
      "NeuralArchitect output shape: torch.Size([1, 10])\n",
      "\n",
      "FLOPs: 1.95M\n",
      "Params: 1.61M\n",
      "Conv2d output shape: torch.Size([1, 16, 28, 28])\n",
      "\n",
      "Layer: Conv2d\n",
      "Input shape: torch.Size([1, 3, 28, 28])\n",
      "Output shape: torch.Size([1, 16, 28, 28])\n",
      "Parameters: 448\n",
      "Weight stats:\n",
      "  Mean: -0.0072\n",
      "  Std: 0.1088\n",
      "Bias stats:\n",
      "  Mean: -0.0412\n",
      "  Std: 0.0895\n",
      "\n",
      "Layer: Conv2d\n",
      "Kernel size: 3\n",
      "Stride: 1\n",
      "Padding: 1\n",
      "Receptive field increase: 2\n",
      "ReLU output shape: torch.Size([1, 16, 28, 28])\n",
      "Linear output shape: torch.Size([1, 128])\n",
      "\n",
      "Layer: Linear\n",
      "Input shape: torch.Size([1, 12544])\n",
      "Output shape: torch.Size([1, 128])\n",
      "Parameters: 1605760\n",
      "Weight stats:\n",
      "  Mean: 0.0000\n",
      "  Std: 0.0052\n",
      "Bias stats:\n",
      "  Mean: -0.0005\n",
      "  Std: 0.0050\n",
      "ReLU output shape: torch.Size([1, 128])\n",
      "Linear output shape: torch.Size([1, 10])\n",
      "\n",
      "Layer: Linear\n",
      "Input shape: torch.Size([1, 128])\n",
      "Output shape: torch.Size([1, 10])\n",
      "Parameters: 1290\n",
      "Weight stats:\n",
      "  Mean: 0.0024\n",
      "  Std: 0.0519\n",
      "Bias stats:\n",
      "  Mean: 0.0015\n",
      "  Std: 0.0489\n",
      "NeuralArchitect output shape: torch.Size([1, 10])\n"
     ]
    }
   ],
   "source": [
    "# Create an instance of the NeuralArchitect\n",
    "network = NeuralArchitect()\n",
    "\n",
    "# Print the network architecture\n",
    "print(network)\n",
    "\n",
    "# Use torchsummary to get a detailed summary of the network\n",
    "summary(network, (3, 28, 28))\n",
    "\n",
    "# Analyse the network parameters\n",
    "total_params = sum(p.numel() for p in network.parameters())\n",
    "trainable_params = sum(p.numel() for p in network.parameters() if p.requires_grad)\n",
    "\n",
    "print(f\"Total parameters: {total_params}\")\n",
    "print(f\"Trainable parameters: {trainable_params}\")\n",
    "\n",
    "# Examine the output shape at each layer\n",
    "sample_input = torch.randn(1, 3, 28, 28)\n",
    "\n",
    "def hook_fn(module, input, output):\n",
    "    print(f\"{module.__class__.__name__} output shape: {output.shape}\")\n",
    "\n",
    "for name, module in network.named_modules():\n",
    "    if not isinstance(module, nn.Sequential):\n",
    "        module.register_forward_hook(hook_fn)\n",
    "\n",
    "_ = network(sample_input)\n",
    "\n",
    "# Analyse the activation distributions\n",
    "activations = {}\n",
    "\n",
    "def activation_hook(name):\n",
    "    def hook(module, input, output):\n",
    "        activations[name] = output.detach()\n",
    "    return hook\n",
    "\n",
    "for name, module in network.named_modules():\n",
    "    if isinstance(module, (nn.ReLU, nn.Linear, nn.Conv2d)):\n",
    "        module.register_forward_hook(activation_hook(name))\n",
    "\n",
    "_ = network(sample_input)\n",
    "\n",
    "for name, activation in activations.items():\n",
    "    print(f\"{name} activation stats:\")\n",
    "    print(f\"  Mean: {activation.mean().item():.4f}\")\n",
    "    print(f\"  Std: {activation.std().item():.4f}\")\n",
    "    print(f\"  Min: {activation.min().item():.4f}\")\n",
    "    print(f\"  Max: {activation.max().item():.4f}\")\n",
    "\n",
    "# Analyse gradient flow\n",
    "network.zero_grad()\n",
    "output = network(sample_input)\n",
    "output.sum().backward()\n",
    "\n",
    "for name, param in network.named_parameters():\n",
    "    if param.requires_grad:\n",
    "        print(f\"{name} gradient stats:\")\n",
    "        print(f\"  Mean: {param.grad.mean().item():.4f}\")\n",
    "        print(f\"  Std: {param.grad.std().item():.4f}\")\n",
    "        print(f\"  Min: {param.grad.min().item():.4f}\")\n",
    "        print(f\"  Max: {param.grad.max().item():.4f}\")\n",
    "\n",
    "# Visualize weight distributions\n",
    "weight_params = [(name, param) for name, param in network.named_parameters() if 'weight' in name]\n",
    "num_weight_layers = len(weight_params)\n",
    "\n",
    "fig, axes = plt.subplots(num_weight_layers, 1, figsize=(12, 4*num_weight_layers))\n",
    "axes = axes.flatten() if num_weight_layers > 1 else [axes]\n",
    "\n",
    "for i, (name, param) in enumerate(weight_params):\n",
    "    ax = axes[i]\n",
    "    ax.hist(param.detach().cpu().numpy().flatten(), bins=50)\n",
    "    ax.set_title(f'{name} distribution')\n",
    "    ax.set_ylabel('Frequency')\n",
    "    ax.set_xlabel('Weight Value')\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.show()\n",
    "\n",
    "# Layer-wise analysis\n",
    "def layer_analysis(module, input, output):\n",
    "    print(f\"\\nLayer: {module.__class__.__name__}\")\n",
    "    print(f\"Input shape: {input[0].shape}\")\n",
    "    print(f\"Output shape: {output.shape}\")\n",
    "    print(f\"Parameters: {sum(p.numel() for p in module.parameters())}\")\n",
    "    if hasattr(module, 'weight'):\n",
    "        print(f\"Weight stats:\")\n",
    "        print(f\"  Mean: {module.weight.mean().item():.4f}\")\n",
    "        print(f\"  Std: {module.weight.std().item():.4f}\")\n",
    "    if hasattr(module, 'bias') and module.bias is not None:\n",
    "        print(f\"Bias stats:\")\n",
    "        print(f\"  Mean: {module.bias.mean().item():.4f}\")\n",
    "        print(f\"  Std: {module.bias.std().item():.4f}\")\n",
    "\n",
    "for name, module in network.named_modules():\n",
    "    if isinstance(module, (nn.Conv2d, nn.Linear)):\n",
    "        module.register_forward_hook(layer_analysis)\n",
    "\n",
    "_ = network(sample_input)\n",
    "\n",
    "# Compute FLOPs and memory usage\n",
    "flops, params = profile(network, inputs=(sample_input,))\n",
    "print(f\"\\nFLOPs: {flops/1e6:.2f}M\")\n",
    "print(f\"Params: {params/1e6:.2f}M\")\n",
    "\n",
    "# Analyse receptive field\n",
    "def receptive_field(module, input, output):\n",
    "    if isinstance(module, nn.Conv2d):\n",
    "        kernel_size = module.kernel_size[0]\n",
    "        stride = module.stride[0]\n",
    "        padding = module.padding[0]\n",
    "        print(f\"\\nLayer: {module.__class__.__name__}\")\n",
    "        print(f\"Kernel size: {kernel_size}\")\n",
    "        print(f\"Stride: {stride}\")\n",
    "        print(f\"Padding: {padding}\")\n",
    "        print(f\"Receptive field increase: {kernel_size - 1}\")\n",
    "\n",
    "for name, module in network.named_modules():\n",
    "    if isinstance(module, nn.Conv2d):\n",
    "        module.register_forward_hook(receptive_field)\n",
    "\n",
    "_ = network(sample_input)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}